The Roomba for Lawns Is Really Pissing Off Astronomers

http://www.wired.com/2015/04/irobot-lawnbot/

by digital55 • 11 years ago

78 116 11 years ago

briandear• 11 years ago

Wow. That article took 8 paragraphs or so before it explained the issue that was irritating astronomers. Terrible writing.

veidr2• 11 years ago

Yep, completely contrary to everything I learned in journalism school.

TL;DR (from the seventh and last paragraph) — "The frequency band proposed for the lawnbot (6240-6740 MHz) is the very same one several enormous radio telescopes operate on."

pc86• 11 years ago

There are 14 paragraphs in the article.

smhg• 11 years ago

Depends on the goal I guess. What you might have been taught: inform your readers.

The writer's goal/task was probably something like "create content" in the broad sense.

My personal opinion: please keep doing what you learned in journalism school :)

e40• 11 years ago

Filling space is how they justify their jobs. The entire article could have been a few small paragraphs.

jiballer• 11 years ago

I'm curious why it needs beacons when we have things like self-driving cars and machine learning? Is it just a matter of cost?

Piskvorrr• 11 years ago

Indeed. Easier on customer service, too: "and are you SURE that you have your beacons positioned correctly?"

bpires• 11 years ago

How are you going to know where someone's property ends? You're not going to equip a lawn mower with a mounted laser range finder and sophisticated image processing capabilities.

sp3322• 11 years ago

Self-driving cars barely exist as prototypes. It's not production-ready tech.

chc• 11 years ago

Don't Tesla already have their autopilot tech released?

fooqux• 11 years ago

Because I don't have a double-yellow line painted between my property and my neighbors? Machine learning is fine and all, but you need something for the machine to learn from. Without some kind of outside influence, my grass looks the same as the grass across the invisible property line.

Not only that, but you need a system to account for wheel slippage, etc. No matter how accurate your compass, how awesome your internal map, you're eventually going to go off course. Best case, your bot mows your neighbor's property. Worst case, it wanders into the busy street, or into your pond. Systems like these beacons or underground wire help ensure this doesn't happen.

Source: I have a half-built robot lawnmower on my workbench.

CPLX3• 11 years ago

we have things like self-driving cars

On the planet I live on we don't have cars that are able to drive themselves.

0xffff23• 11 years ago

Which planet would that be? I see them nearly every day. They certainly aren't perfect, but I think it's reasonable to say that we have working prototypes of self-driving cars.

CaveTech• 11 years ago

How do you machine learn your property boundaries? It's an abstract concept of personal ownership.

wpietri• 11 years ago

I have pal who did sysadmin work for a radio telescope; they went to amazing lengths to avoid interference. The details escape me now (hopefully some of those folks will chime in with their stories), but I remember being impressed about how thoughtful and serious they were. So between the radio astronomers and corporate lawyers, I'm inclined to believe the astronomers.

hellweaver666• 11 years ago

Why can't we just manually "walk" the mower around the border so it knows where it's boundries are?

detaro• 11 years ago

Because the hard part is figuring out where it actually is relative to these borders when it runs around later. If the border is actively detectable (beacons, buried wire, ...) that becomes easier.

Zikes• 11 years ago

It would still need one or more fixed points of reference to keep track of its position. It wouldn't be able to reliably tell where it currently is in relation to where it's been.

jarman• 11 years ago

Problem is probably not in boundary setting, but in bot positioning. Anything fancier than beacons would raise production or support costs too much for mass consumer product

drzaiusapelord3• 11 years ago

The "internet of things" is just going to make this worse and there's really no fighting it. My question is considering how many billions are pissed away tossing men and women into the ISS for reasons not apparent to me, why don't we have a radio telescope on the far side of the moon yet? I keep hearing about all these manned missions, COTS money handed to SpaceX, etc but the one thing that would be a gamechanger for radio astronony is not even on anyone's radar.

I'd love to see the next POTUS make this a priority. Yeah, Orion and the SLS are impressive, but they're mostly fighting yesterday's battles - pretty much a souped up Apollo for similar big feel-good missions where men and women go into LEO/moon/asteroids/maybe-mars for a very short time and fly right back. Why can't we also fund a farside telescope that could work for decades 24/7 providing space science advantages we simply cannot get any other way?

g-clef2• 11 years ago

The moon is a pretty bad place for any long-term base...lunar dust is really hard on equipment. The l4/l5 points, though, are very interesting and there's a bunch of projects putting probes at the various l4/l5 points in the solar system.

http://en.wikipedia.org/wiki/List_of_objects_at_Lagrangian_p...

titzer• 11 years ago

There's no atmosphere, so its unlikely Lunar dust would contaminate any equipment. It settles immediately after landing and doesn't blow around. Minimal shielded for touchdown would be all that's necessary.

drzaiusapelord3• 11 years ago

Perhaps I wasn't clear. This would be an unmanned telescope. We're talking a one time landing of a piece of equipment. Dust will be a non-issue after the initial landing.

TorKlingberg• 11 years ago

"Internet of things" devices mostly use cellular, WiFi or Bluetooth, and those frequencies are already in heavy use so I doubt it will make much difference for astronomers.

About the far side of the moon, there is little reason to put a telescope there. It is much easier to put one in orbit around the earch, sun or an earth-sun Lagrange point.

As for manned space exploration versus robotic, I think that is a debate for an other day.

drzaiusapelord3• 11 years ago

Those other locations still be susceptible to stray radio transmissions, especially an Earth orbit which would also have issues with Earth's ionosphere. All that moon rock is going to be a great radio insulator.

I also imagine there are cost savings by having it sit on the surface of a moon than being its own satellite. No need for worry about orbit corrections or strict weight limitations. The Tsiolkovskiy crater is pretty flat and would make a good candidate for a telescope which would have long antenna arrays laying on the surface.

ohashi• 11 years ago

So why can't they just use a different frequency?

LandoCalrissian• 11 years ago

Kind of my question. Why can't they use one of the standard open frequencies like 433mhz? I assume there is some reason, maybe broadcast distance and strength for bigger yards?

cevanwells• 11 years ago

Exactly. What makes the 6Ghz spectrum so attractive to iRobot? If they are asking the FCC for an exception anyway, isn't there another frequency they could use with the same method of location finding?

userbinator• 11 years ago

I'm not sure why they're so obstinate about using a very-high-frequency band of RF for this. Is GPS not accurate enough? How about a more mechanical method, or near-field communications?

the only way to get them to work is to dig a trench along the perimeter of a lawn and install a wire that creates the electronic fence needed to ensure the automaton don’t wander beyond the property line.

If they really want to implement that method, shopping cart locking systems have done it before, and they don't need 6GHz+ RF:

http://www.wired.com/2007/07/hack-remote-sho/

pmorici• 11 years ago

You mis read it, they want to _avoid_ having a buried wire system. They are instead proposing a stake system which is what they are requesting a waiver for.

rwmj• 11 years ago

Tomorrow's World, 1979: https://www.youtube.com/watch?v=Ade1JYpscl0

I wonder if the real reason is that they can't get a patent for simpler methods because it's all been done decades ago?

TeMPOraL4• 11 years ago

I so wouldn't be surprised if that would be the actual reason.

cwyers• 11 years ago

Is GPS not accurate enough?

It really isn't.

teamonkey• 11 years ago

Or optical, even? Once you've accepted that there will need to be a stake in the ground, you'd already eliminate a lot of the line-of-sight problems.

pjc502• 11 years ago

GPS accuracy on its own is nowhere near good enough. Differential GPS can give you 1cm accuracy: http://www.blackroc-technology.com/technologies/gnss/dgps/

0xffff23• 11 years ago

It seems like installing the base station for differential GPS would be easier than installing the electronic fencing descried in the article.

hackmiester• 11 years ago

Your quote is of exactly the method they do not want to use.

mindslight• 11 years ago

I've got to believe there's good reasoning from iRobot, even though none of the articles I've seen have covered it (tsk tsk IEEE Spectrum). Perhaps wanting to bury the transceivers rules out 2.4GHz and 5GHz, while needing tight directionality and small antennas rules out the lower stuff?

Having said that, radio interference is taken pretty seriously (eg http://en.wikipedia.org/wiki/United_States_National_Radio_Qu...). The free-for-all that we're familiar with is really just a few tiny consumer bands.

DGPS seems like a better idea (even easier to install a virtual border by manually driving the robot around the edge), but would probably require some above-surface beacons. I'd imagine iRobot is trying to cater to the type of people who think grass needs to be mowed every week and wouldn't want even a few beacons sticking up.

Still I'm sure there's another way. And if iRobot wants to win in even just the court of public opinion, they have a steep argument to make.

ryandrake• 11 years ago

GPS isn't enough. A lot depends on the receiver, antenna, how many satellites are in view, and what's going on with the atmosphere, but unaugmented GPS is generally considered accurate to about 15 meters. Not good enough for a lawn unless you're in a rural area or your neighbors don't care if you cross the property line.

GPS in your car only appears to be more accurate than that because they're typically "snapping" the GPS signal to an underlying road map.

rbarris• 11 years ago

I would wager they are using UWB for position tracking, as seen in the DecaWave devices, which can operate in the 6GHz spectrum. No GPS required, but you do need multiple nodes to range against.

http://www.decawave.com

JoeAltmaier• 11 years ago

It comes down to some engineer choosing a certain chip and antennae. iRobot, come on, choose a different vendor and end this silliness.

josefresco3• 11 years ago

I'm all for robotics, but the thought of an unattended machine whirling around my yard with spinning blades (of death) worries me. Granted, my "lawn guy" races around my yard already like a NASCAR driver on crack, but part of me wonders what sort of mishaps this robotic vacuum might get itself into. Small animals, toys, sprinkler heads etc.

It's one thing if your vacuum sucks up a kid's toy, or harasses your cat. Imagine the potential liability if your robotic lawn mower left your yard and entered your neighbors.

ptaipale3• 11 years ago

Stiga has been making robot mowers for quite a while, and I haven't heard of drunken Swedes who would have been hacked to death by their mowers after passing out on their lawn.

Apparently animals will run away and sprinkler heads can be marked with perimeter wire so that the gadget avoids them.

http://www.stiga.com/products/stiga_en/robot-mowers.html

josefresco3• 11 years ago

I have sprinkler heads all over my lawn, if the lawn mower avoided them I'd have patches of long grass. My sprinkler heads will routinely get stuck in the "up" position due to sand/dirt that gets stuck in the sliding mechanism.

This actually seems solvable by robotics, but it's not a viable solution for beacons or markers as they are meant to be "passed over" by lawn mowers.

jkmcf• 11 years ago

Right now, Stiga is wondering WTF is going on with their website.

melling• 11 years ago

This is from a TechCrunch article published 3 days ago:

http://techcrunch.com/2015/04/13/irobot-is-working-on-a-robo...

"iRobot: “Use of the iRobotRLM [robot lawn mower] will increase lawn mower safety. An estimated 1,517 lethal accidents occurred with lawn mowers through the years 1997 to 2010. It is reasonable to assume that many of these injuries and deaths would not occur if consumers used a robotic mower. More than 17 million gallons of fuel, mostly gasoline, are spilled each year while refueling lawn equipment. Abattery poweredRLM will reduce emissions, gasoline spills, fires and other such accidents.”NRAO: “iRobot cited multiple statistics of grim accidents and spilt gasoline to assert the public benefit of approving its wireless robotic lawn mowers. However, there is already a competitive market for robotic lawn mowers using wire loops [buried edge wire], which has somehow failed to stanch the stream of ghastly accidents and spilt gasoline that iRobot associates with the mundane practice of lawn-mowing.”"

josefresco3• 11 years ago

If you actually read the paragraph you'll find some (contradictory) issues with the NRAO statement (update: see below). Keep in mind we're talking about a statement made by astronomers about lawn mower safety.

"An estimated 1,517 lethal accidents occurred with lawn mowers through the years 1997 to 2010."

No data on the "1,517 lethal accidents" to prove that they were caused by fires or anything that a robotic vacuum solves (although I'd be happy to be sent a link proving this). So it's not "reasonable" to assume that these injuries will be reduced - the icing on the cake is the last sentence which essentially contradicts the first...

"there is already a competitive market for robotic lawn mowers using wire loops [buried edge wire], which has somehow failed to stanch the stream of ghastly accidents and spilt gasoline that iRobot associates with the mundane practice of lawn-mowing"

Edit: Just realized the fist statement (claiming safety) is by iRoomba, the second statement (disputing this) is by the NRAO.

So iRoomba (the one selling the product) claims it will reduce vacuums. The NRAO pointed out that despite a "competitive market" for robotic lawn mowers injuries have not been reduced.

Edit 2: Data! http://www.ncbi.nlm.nih.gov/pubmed/16361919

Brakenshire2• 11 years ago

You would assume, though, that most of those accidents are people using or maintainong the lawn mower. It's quite different if it damages someone else entirely. 'Robotic lawnmower runs over neighbour's child' would be a massive story even if it only happened once.

uptown• 11 years ago

Is this the same technology used by invisible fences intended to contain dogs in a yard?

cpwright2• 11 years ago

Yes, but they are arguing that using the invisible fence wire is too costly/difficult for a lawnmower. I can see that argument. A basic consumer lawnmower is probably $200-600 right now. An installed pet fence is quite a bit more.

Also, installing a pet fence, e.g. across driveways and paths can involve asphalt/concrete cutting, and a bunch of trenching. That is the kind of thing that even if you're willing to pay may have you say that it is just not worth it, compared to sticking some stakes in the ground.

The market is interesting here; you have to have someone who will spend enough on a fancy lawnmower, but not so much that they would outsource it to a landscaper.

kstenerud4• 11 years ago

Why don't they just use the WIFI spectrum?

Have legitimate messages so that it doesn't mess up existing systems, and send periodic pings whose radio signals the lawnmower can track. As an added bonus, you'd have a full communications network to talk to it with.

gee_totes• 11 years ago

How will the lawnmower know where it is with just WiFi?

kstenerud4• 11 years ago

The same way it would know where it is with the 6240-6740 MHz beacons.

snowwrestler3• 11 years ago

By analyzing differentials in the signals from each beacon, which is how WiFi positioning systems work now.

http://en.wikipedia.org/wiki/Wi-Fi_positioning_system

Anyway, WiFi actually operates in unlicensed spectrum, not in WiFi-specific spectrum, so continuous carrier waves would be legal too.

jfabre• 11 years ago

Am I the only one who thinks that iRobot is actually doing a very poor job with the Roomba cleaning algorithm?

Why can't we have a robot that maps the surface and then cleans accordingly? I always get irritated when I look at my Roomba doing it's job.

chiph• 11 years ago

The Neato vacuum does a back & forth pattern. But after living with a Roomba for a while, I've grown to appreciate their mostly-random pattern - it (eventually) makes sure that the nap of the carpet gets brushed in different directions, so your carpet doesn't look like the greens at the local golf course, with stripes in it.

t0mbstone• 11 years ago

Look at the Neato. I switched from Roomba to Neato and never looked back. It actually vacuums in an intelligent manner.

Luc• 11 years ago

Because you need more processing power and sensors to first build the map, and then know where you are on the map.

akramhussein2• 11 years ago

Check out the upcoming Dyson 360 Eye - http://www.dyson360eye.com/

Disclaimer: I worked on it.

Edit: To answer your question directly, you have 3 choices:

1. Don't map (or somewhat guess it) - Roomba.

2. Map first, then clean - Neato.

3. Map as you clean - Dyson 360 Eye.

This is obviously a crude explanation but they get harder going down the list. If you are interested to know more, read up on SLAM (Simultaneous Localisation and Mapping) algorithms.

robterrell2• 11 years ago

How is it better than Roomba and Neato -- what's the cleaning algorithm? Does it map the space? etc.

Edit: thanks, you answered while I was asking. That's a very interesting use case for SLAM. Does it re-map every use (i.e. today there's toys on the floor, but tomorrow there's not)? Does it store the point cloud or convert it to something more manageable? Are you licensing someone's SLAM library or did you roll your own?

berberous• 11 years ago

It does a pretty good job though:

http://www.designboom.com/weblog/images/images_2/jenny/roomb...

I tried a Neato but it was much more prone to getting stuck (every single time actually). Obviously that will vary depending on your furniture.

IndianAstronaut• 11 years ago

This is actually better done with a robot team or swarm. There are several 'frontier' algorithms which use basic particle swarm or metaheuristic algorithms to rapidly move across a landscape or frontier.

snowwrestler3• 11 years ago

What's better, a brute force algorithm that is cheap and simple but slow, or a highly optimized system that is expensive and complex and fast?

The answer is of course "it depends." Certainly the Roomba could be smarter, faster, and more efficient, but the question is how much more are customers willing to pay for those.

wppick• 11 years ago

Can we just get rid of grass lawns already... replace it with something that looks good and doesn't need water and weekly cutting

mrfusion2• 11 years ago

Agreed. I think it's so amazing we actually make up work for ourselves.

oh_sigh8• 11 years ago

It's not "making up work". People value a lawn, and put in work to maintain it. I find it funny that so many people on hacker news don't understand this concept. I'm sure a lot of people here would love to live in a concrete cube and drink soylent all day long while plugged into the internet on their oculus VR, but it's shocking to me that you can't comprehend how other people can value different things from yourself.

at-fates-hands2• 11 years ago

Not sure if this was sarcasm or not, but having a lot of landscaping architecture friends it's an interesting idea.

In places where grass is hard to grow (Southwestern US) you see a lot of rock varieties and other materials that they substitute for grass and it looks great.

Just Google, "las vegas yards" and look at the image search.

The last place I lived, my neighbor was tired of his grass in his bakyard (about a 1/4 acre) so he created a huge concrete terrace in his back yard, along with a large rubber hard court for basketball and other activities for his and the neighborhood kids. He still kept his small front yard grass, but the entire backyard was now completely utilitarian.

rmxt• 11 years ago

Yay for (sub?)urban heat island growth! :(

Not sure where the last place you lived was, but I hope that if it were the desert, that there's some sort of aesthetically pleasing option that exists between a water hungry green grass lawn, and an asphalt or impervious concrete surface. Even if it's not the desert, permeable paving sounds like a win-win solution for aesthetics, functionality, and environmental impact.

https://en.wikipedia.org/wiki/Permeable_paving

patorjk• 11 years ago

When I lived in the New Mexico in the early 90's people started overhauling their grass yards with rock landscaping (with a few plants dotted in). The results tended to look great. I'm surprised this hasn't taken off in more places across the country.

abetaha• 11 years ago

In Arizona and other water starved states they use artificial grass in their lawns, and some of them look reasonable. I am starting to see the trend in the bay area as well.

oh_sigh8• 11 years ago

Instead of not having grass yards, maybe people should just go the full mile and not live in areas completely inhospitable to human life?

ptaipale3• 11 years ago

Yeah, an Astroturf home lawn is what we want.

Seriously, lawns, like other parts of gardens, are something that people do for hobby - they like the work. Yes, it is somewhat contradictory that then people also want to avoid the chores of mowing. A rock garden may be better but anything will need maintenance work to look neat.

oh_sigh8• 11 years ago

Water is not scarce in a lot of areas.

Also, you make it sound like the government is paying for people to cut their yards. People do it themselves out of their own free will.

milesokeefe• 11 years ago

People do it themselves out of their own free will.

In many places in the US, you are legally required to keep your lawn below a certain height, and not just in HOA neighborhoods, it is often enforced by the city.

Brakenshire2• 11 years ago

If the alternative is some more appropriate local plant or plants, then great. If the alternative is concrete or astroturf, please don't.

The trend of people paving over their front gardens for car parking is causing a lot of flooding problems in the UK (because water runs straight off concrete, and isn't absorbed). It also makes the general environment much less pleasant and human if everything surrounding roads is paved as far as the eye can see, rather than filled with gardens, whether they're wild or highly tended. You end up with these grim concrete tunnels of development where no one wants to spend their time. That in turn damages pedestrian culture, and the health and social benefits which go with it.

dwb• 11 years ago

Mind your own business! http://www.bbc.co.uk/gardening/plants/plant_finder/plant_pag...

kalleboo• 11 years ago

Husqvarna have been making robotic lawnmowers since 1995, do they work differently? http://www.husqvarna.com/us/products/robotic-mowers/husqvarn...

seszett• 11 years ago

As the article says, you have to bury wire along the perimeter that is to be mowed by the robot, that's how it recognizes its limits.

ptaipale3• 11 years ago

Like Stiga and others, they are based on having a perimeter wire on the ground.

grandalf• 11 years ago

It's nice that RFI is being taken seriously. In many, many cases it is not: Noisy utility transformers, plasma TVs, poorly designed wall-wart supplies, touch lamps, and broadband over power lines, just to mention a few troublesome sources.

mrfusion2• 11 years ago

Would GPS augmentation not work for this case? Don't airport runways use a system like that?

mhb• 11 years ago

It would be nice to crack down on the lawnmower and, especially, leaf blower emissions in the 40Hz - 10KHz band. Where's the outrage?

rebootthesystem• 11 years ago

I don't think lawn bots will take off. I live in a nice neighborhood. I am 100% certain that a $500 to $3,000 lawn bot would be stolen within a month.

Also, it is far cheaper and more efficient to pay a gardener for lawn maintenance. They take care of far more than just mowing the lawn.

kleinsch• 11 years ago

Not an expert, but isn't this the exact problem Bluetooth LE was designed to solve?

thaumasiotes• 11 years ago

In a later response, iRobot added that NRAO observatories usually are surrounded by desert or forests, not environments where residential lawn equipment is used—a claim the NRAO called “silly.”

I don't get what's silly about this. Astronomers are already well known for hating nearby human residences, because they give off visible light that makes it difficult to see the sky. Radio communications are the exact same thing at a different frequency.

But we've got a strong system going of "astronomers can suck it, and if they want to look at the sky without our light getting in the way they can live up on a remote mountaintop where nobody else ever goes". And that's what they do, because imagine if someone could build a telescope near your house and then get laws passed saying you were prohibited from using electric lighting... ever. We like being able to see at night. Maybe we also like having automatic lawn maintenance. Why should astronomers get to stop us?

TeMPOraL4• 11 years ago

Because RF spectrum is big and there is really nothing forcing iRobot to use the same band radio astronomers need (who need that particular band because physics) - they're probably either too lazy to pick a different radio chip for their product or looking for a clever way to avoid the good solutions we have since like 70s, so they could patent something here.

merrua• 11 years ago

iRobot really should use a different frequency. Radio Telescopes should have the prior claim and are more useful and valuable.

eru2• 11 years ago

They even refrained from the clickbait of mentioning methanol in the headline:

"... pissing off astronomers looking for a drink"

tzs• 11 years ago

iRobot's waiver request: http://apps.fcc.gov/ecfs/document/view?id=60001028220

NRAO comment on waiver request: http://apps.fcc.gov/ecfs/document/view?id=60001039394

iRobot reply to comment: http://apps.fcc.gov/ecfs/document/view?id=60001041441

NRAO reply to reply to comment: http://apps.fcc.gov/ecfs/document/view?id=60001042025

I see nothing in the iRobot filings that indicate why they cannot operate in the 5.725-5.875 GHz ISM band. Their system is based on simple time of flight measurements of the signal, and that should work just as well at 5.7-5.9 GHz as it would at 6.2-6.7 GHz.

Qworg• 11 years ago

I was wondering when this would come up. I led a team building autonomous lawnmowers for a major outdoor power equipment company.

To answer some questions:

Autonomous lawnmowers are remarkably safe - far safer than standard lawnmowers by a wide margin.

iRobot's potential solution isn't a new one by any means - but the requirement of getting FCC approval for transmission is a rough requirement to overcome.

GPS isn't sufficient and DGPS can't fix multipath problems. Bosch's lawnmower had GPS "straight line" algorithms in it and it had to be pulled because it was destroying people's lawns.

Decawave is an interesting technology, but not what they're using.

Fully autonomous lawnmowers have been around since the 1950's and haven't changed overmuch: http://file.vintageadbrowser.com/tg2i5rpfzmy0yj.jpg

The Roomba for Lawns Is Really Pissing Off Astronomers

Related Stories